Shape-Preserving Prediction for Stationary Functional Time Series

摘要: This article presents a novel method for prediction of stationary functional time series, trajectories sharing similar pattern with phase variability. Existing methodologies series only consider amplitude To overcome this limitation, we develop that incorporates One major advantage our proposed is the ability to preserve by treating as shape objects defined in quotient space respect warping and jointly modeling estimating Moreover, does not involve unnatural transformations can be easily implemented using existing software. The asymptotic properties least squares estimator are studied. effectiveness illustrated simulation study real data analysis on annual ocean surface temperatures. It shown SP (shape-preserving) captures common better than method, while providing competitive accuracy.